Management of oedematous malnutrition in

infants and children aged >6 months:

a systematic review of the evidence

Dominique Roberfroid1,2
Naïma Hammami1,3
Pankti Mehta1
Carl Lachat1,4
Roos Verstraeten1,4
Zita Weise Prinzo5
Lieven Huybregts1,4
Patrick Kolsteren1,4

October 2013

1
Woman & Child Health Research Centre, Institute of Tropical Medicine, Antwerp, Belgium
2
Belgian Health Care Knowledge Centre, Brussels, Belgium
3
Institute of Public Health, Brussels, Belgium
4
Department of Food Safety and Food Quality, Ghent University, Belgium
5
Department for Health and Development, WHO HQ, Geneva

Corresponding author: Dominique Roberfroid, Woman & Child Health Research Centre,
Institute of Tropical Medicine, 155 Nationalestraat, 2000 Antwerp, Belgium,
droberfroid@itg.be

This work was commissioned by the World Health Organization.

Authors declare that they have no conflict of interest.

1
Abbreviations

CI confidence interval
CMAM Community Management of Acute Malnutrition
F-75 therapeutic milk used in stabilization phase of the treatment of SAM
HIV human immunodeficiency syndrome
IQR interquartile range
MAM moderate acute malnutrition
MUAC mid-upper arm circumference
RCT randomized controlled trial
RUTF ready-to-use-therapeutic-food
SAM severe acute malnutrition
SD standard deviation
W/A weight-for-age
W/H weight-for-height
WHO World Health Organization
WHZ weight-for-height z-score

Measurements
d day
cm centimetre
g gram
kcal kilocalorie
kg kilogram
mg megajoule
mm milimetre
n number

2
Abstract
Background
Children presenting mild or moderate nutritional oedema are usually treated in outpatient
care. However, it is still unclear whether such children presenting require special care and if
they can be treated indifferently in inpatient or outpatient care.

Methods
We systematically searched Medline, Embase, the Cochrane Library and DARE for evidence
on mortality, recovery, treatment duration, weight gain, cost-effectiveness and adverse events
in children with nutritional oedema treated in outpatient care as compared to treatment in
inpatient care or to treatment of marasmus in outpatient care. We complied with the PRISMA
statement for reporting of systematic review.

Results
Eight studies met the inclusion criteria, among which six observational studies and two
controlled trials. These two trials assessed outcomes of outpatient care after one week of
inpatient stabilization, i.e. not direct outpatient care. The first trial in Malawi reported better
results in outpatients than in inpatients (recovery: 72% vs 49%) probably related to more
infections in inpatient care. The second trial in Bangladesh reported the reverse (recovery:
67% vs 86%) but no food supplement was provide to outpatients. The recovery rates for
outpatients ranged across studies from 69% to 96%. These rates were consistently higher than
for those children with marasmus (range 33–81%) or marasmic kwashiorkor (range 45–71%),
and mortality rates were consistently lower. No evidence on cost-effectiveness or adverse
events was available. Overall, the quality of the evidence was low.

Conclusion
Oedematous malnutrition could plausibly be treated effectively in outpatient service.
However, the quality of evidence was low and further good quality studies in various settings
are required before conclusive guidance can be generated.

3
Background and objective
Along with a low mid-upper arm circumference (MUAC <115 mm) or a low weight-for-
height <-3 z-score (WHZ; WHO 2006 growth standards), bilateral pitting oedema, also called
nutritional oedema, is an independent indicator of severe acute malnutrition (SAM) in 6–59
month old infants (1). Three grades of severity are distinguished: oedema + is defined as a
mild oedema on both feet/ankles; oedema ++ is a moderate oedema on feet and lower legs,
hands or lower arms; and oedema +++ is severe oedema generalized to feet, legs, hands, arms
and face (2,3).

The standard recommendation for the management of SAM has been inpatient care during the
stabilization and rehabilitation phases (4). However, inpatient treatment is resource intensive
and requires many skilled and motivated staff, which is often lacking in low- and middle-
income countries where SAM is common (3). Moreover, studies have shown that children
with SAM could be adequately treated at home (5–8). These results have led to a revised
classification that categorizes acute malnutrition in complicated cases, which require inpatient
care in the first stage of treatment, and uncomplicated cases that can be treated in the
outpatient setting or at the community level (3,9). In the absence of severe illness and if an
adequate appetite is maintained, SAM is considered uncomplicated. Regarding nutritional
oedema, only children with oedema +++ or oedema +/++ combined with either a low MUAC,
poor appetite or a severe illness should be hospitalized in the first stage of the treatment (3).
These recommendations apply in many countries. Among 33 national guidelines on
Community Management of Acute Malnutrition (CMAM) collected by the Emergency
Nutrition Network in November 2011 (http://cmamconference2011.org/latest-conference-
updates/), almost all admitted uncomplicated oedema +/++ to outpatient care if there was
good appetite (Annex 1). There were only six countries in which nutritional oedema,
regardless of its severity, was systematically referred for inpatient treatment (Angola, Mali,
Mozambique, Niger, Tanzania and Zanzibar) and another two where only oedema + was
admitted to outpatient care (Côte Ivoire and Kenya).

A recent joint statement by the World Health Organization (WHO), the United Nations
Children’s Fund and the World Food Programme also advocates treating children with SAM
at the community level whenever possible (10). However, this guidance does not differentiate
grades of oedema and states generically that “children can be transferred from facility to
community-based care when their oedema is reduced” (10). Actually, it is still unclear
whether children presenting nutritional oedema +/++ require special care and if they can be
treated indifferently in inpatient or outpatient care, i.e. if the critical outcomes of survival and
recovery are similar in both settings. Until now, the evidence has not been reviewed in a
systematic way.

We, therefore, aimed at summarizing the evidence on the effectiveness and safety of
managing children >6 months with uncomplicated oedematous SAM grade +/++ in outpatient
care. This assessment is done by comparison with results obtained in the in-patient treatment
of similar cases or in the out-patient treatment of uncomplicated SAM children without
oedema (low WHZ and/or low MUAC) (Table 1). The results of this review will inform
future policies on a safe and effective management of oedematous malnutrition.

Table 1
Research question (SPICE criteria)

Setting Low- and middle-income countries, primary health

care
Perspective/Population Children >6 months with oedematous SAM +/++, and
good appetite or no medical complications

4
 Intervention Community-based therapeutic care – outpatient
management
Comparison - Uncomplicated nutritional oedema +/++ treated in
inpatient care or health service-based care
- SAM cases without oedema treated in outpatient
care
Evaluation (1) short-term mortality
(2) recovery rate
(3) time to recover
(4) weight, length, and MUAC gain
(5) use of resources, costs
(6) adverse effects

2. Methodology
2.1 Sources and key words for literature review
We searched Medline via PubMed (http://www.ncbi.nlm.nih.gov), Embase
(http://www.embase.com), the Cochrane Library (www.cochranelibrary.com) and DARE
(http://www.crd.york.ac.uk). We also searched the Emergency Network website
(http://ennonline.net/) for additional field reports, and screened the bibliography of included
studies.

We applied the following search strategy:

#1. (((Nutrition Disorders[MeSH] OR Growth disorders[MeSH] OR malnutrition[MeSH] OR

wast*) AND (oedema OR edema)) OR "Kwashiorkor"[Mesh] )
#2. (child[Mesh] OR infant[Mesh])
#3. ((Child Nutrition Disorders[MeSH] OR Infant Nutrition Disorders[MeSH]) AND edema)
#4. (#1 AND #2) OR #3
#5. (treatment OR rehabilitation OR management OR hospitalization OR community health
services)
#6. #4 AND #5
#7. (editorial [pt] OR letter [pt] OR comment [pt])
#8. #6 NOT #7

The search strategy was adapted to fit each individual database (Annex 2). We searched for
studies published up to 08 July 2013. We contacted the authors whenever additional data or
clarifications on the published studies were needed. All publications were integrated into
Reference manager version 12 (ISI ResearchSoft, Thomson Reuters).

2.2 Criteria of inclusion and exclusion

We included all original studies combining the three following criteria: (i) observational or
interventional studies reporting on the outpatient treatment of children >6 months with
nutritional oedema. Studies including children with and without oedema were retained if the
subpopulation with oedema represented ≥80% of the study population or if the results on
children with oedema were reported separately; (ii) studies reporting on any of the following
treatment outcomes: mortality; nutritional recovery; weight/length/MUAC gain; adverse
effects; and cost or cost-effectiveness of treatment; and (iii) studies carried out in low- and
middle-income countries, as defined by the World Bank. We excluded editorials, letters or
comments; studies including only children older than 60 months of age; and studies in
languages other than English, French and Spanish.

5
2.3 Review process
We first screened titles and abstracts of identified references based on the inclusion criteria.
Papers for which the title and/or abstract indicated that inclusion criteria could be fulfilled
were considered and read in full. Two researchers applied the selection process
independently. In case of discrepancies, a third reviewer was consulted and the decision to
include or exclude the paper was reached by consensus. Studies excluded at this second stage
are reported, as well as the reason for their exclusion. Data from included studies were
extracted in predefined standard tables by one reviewer (NH) and cross-checked by a second
one (DR).

2.4 Quality appraisal

The quality appraisal for the selected studies was derived from the checklists of the Scottish
Intercollegiate Guidelines Network (http://www.sign.ac.uk/methodology/checklists.html). We
adjusted the existing checklist for cohort studies to fit case series, as this was the main design
of included studies. The selected quality criteria were: (i) adequate sampling of subjects (i.e.
study not performed on a very specific subpopulation with limited external validity); (ii)
management of missing values (% incomplete data, % of loss-to-follow-up, potential effect of
missing data on results assessed or discussed); (iii) quality of measurement (training of
assessors described, double measurements of anthropometry parameters, quality control
procedures, assessment of accuracy); and (iv) appropriate statistical analysis. Each item was
rated independently by two reviewers as: adequately addressed; moderately addressed; poorly
addressed; or not reported. Disagreement between reviewers was resolved by discussion until
a consensus was reached. In cases of a disagreement, a third reviewer was consulted. We
followed the PRISMA statement for good reporting of systematic reviews
(http://www.prisma-statement.org/).

2.5 Appraisal of program performance

We compared the results reported in the studies with the international standards for good care
of malnourished children (11). According to the Sphere guidelines, the main outcome
indicators for success of management of severely malnourished children are mortality rates
<10%, recovery rates >75%, a recovery time of 30–40 days, default rates <15% and weight
gain > 8 g/kg/day. For fatality rate, we applied the method of Prudhon et al. to calculate the
predicted case fatality rate and compare it to the actual case fatality rate as reported in the
studies (12).

3. Results
We retrieved 999 references from the electronic databases of which 10 papers, reporting on 8
original studies, were included (Figure 1). Ashworth et al. 1997 (13) reported in more detail
on the cost-effectiveness of the controlled trial published in 1994 by Khanum et al. (14).
Ciliberto et al. 2006 (6) reported on children with oedematous malnutrition treated in
outpatient care whose outcomes had already been presented with a 75% overlap in a previous
publication (15) (personal communication of Mark Manary). Annex 3 gives an overview of
excluded studies and reasons for exclusion.

6
Figure 1: Flow chart of review

Records identified through database

searching (n=1135):
Medline (n=777), Embase (n= 324),
Cochrane (n=32), DARE (n=2)

Duplicates (n=136)

Records after duplicates

removed (n=999)

Records excluded based

References manually
on title and abstract
selected (n=11)*
(n=974)

Full text not fulfilling

Full-text articles assessed for
the inclusion criteria
eligibility (n=36)
(n=26)
- No data on oedema
cases (n= 13)
- No eodema cases in
outpatient care (n=5)
- Selected outcomes not
reported (n = 5)
Articles included in the review - Different selection
(n=10), reporting on 8 studies criteria for outpatients
(n=1)
- Other content (n=2)

* References found in the reference list of selected papers. None was retained for this review.

3.1 Characteristics of studies

None of the eight included studies was a randomized controlled trial (RCT) comparing the
outcomes of children with uncomplicated oedema treated exclusively in outpatient vs
inpatient setting (Table 2). Three controlled trials provided useful information, although their
focus differed somewhat from our research question. An RCT with sequential allocation
compared the outcomes of SAM children (98% with oedema) treated in inpatient, day care or
home care after one week of day care in Bangladesh (13,14). Another non-RCT compared the
outcomes of children with SAM (81% oedema) who received inpatient vs home-based care
for the second phase of treatment for childhood malnutrition (6). Finally, an RCT compared
the outcomes of two types of ready-to-use-therapeutic-food (RUTF) during home-based care
of children with SAM (78% with nutritional oedema) (16). Five observational studies

7
reported on the outcomes of oedematous malnutrition (5,17–20), among which four treated
children exclusively in an outpatient setting. Therefore, evidence on treatment of oedematous
malnutrition treated in exclusive outpatient settings originated from five observational studies
– although the study by Oakley et al. (16) was an RCT, it was observational in essence
regarding the research question addressed in this review. Three additional studies provided
evidence on the outcomes of oedematous malnutrition after one week of inpatient
stabilization, two of which were RCTs aimed at comparing outcomes obtained with outpatient
vs inpatient treatment (14,15).

The gradation +/++/+++ was not used in any of the included studies. In three studies, oedema
was graded as severe according to the depth of imprint that remains when the thumb is
removed from the foot dorsum: >0.5 cm (6,15) or >1.0 cm (5,17). Children with severe
oedema were excluded from these studies. No oedema gradation was defined in the remaining
five studies. It is thus assumed that cases of severe oedema were included although the
proportion was not reported (3,14,16,18,20).

Five studies included a proportion of children who combined wasting and bilateral oedema,
but reported results separately for marasmus, kwashiorkor or marasmic kwashiorkor treated
as outpatients (6,16,18–20) (Table 5). Marasmic kwashiorkor was defined as the presence of
bilateral pitting oedema and a WHZ <-3, except in one study in which wasted children were
defined as having a WHZ <-2 (6).

The population of children ranged from 6–60 months, but two studies excluded children
under 10–12 months (6,14,15) and another included children until 120 months (3). The eight
studies recruited a total of 9616 severely malnourished children, from which the majority
(81%) had oedema; 75% of oedema cases were treated solely in outpatient. Nearly all studies
(7/8) took place in African countries and six of eight studies occurred in Malawi.

3.2 Outcomes of intervention

Tables 3–5 summarize the outcomes of the therapeutic nutritional interventions.

3.2.1 Recovery rate

The recovery rates for children with oedema treated in outpatient care ranged across studies
from 69% to 96% (Table 3). These rates were consistently higher than for those children with
marasmus (range 33–81%) or marasmic kwashiorkor (range 45–71%).

Two studies compared outcomes in children treated in outpatient care after one week of
stabilization in inpatient setting vs full inpatient care (Table 4). In Malawi, the recovery rate
of severely malnourished children (81% of whom presented oedema) managed in outpatient
care during the second phase of treatment vs inpatient was, respectively, 72% and 49% at
eight weeks (15). The proportion of children with oedema was equal in each group, but their
specific recovery rates were not reported. However, the authors reported that in a subgroup of
171 children with oedema who were treated exclusively in outpatient management the
recovery rate was 89% (6). Another study in Bangladesh compared the recovery rate of SAM
children (98% of whom presented with oedema) in inpatient care vs day care vs home care
after one week of day care and found, respectively, 87%, 67% and 75% (14). At the start of
home care, 80% of children still had oedema and all recovered (personal communication of
the author). No food supplements were distributed in this latter group.

3.2.2 Mortality rate

The mortality rates for children with nutritional oedema treated in outpatient care reported in
six studies ranged from 1% to 4%, but peaked at 16% in a seventh study (18) (Table 3). One
study in Malawi found a mortality rate of 16% in children with kwashiorkor (18). In that

8
study, 70% of mortality occurred during inpatient care, of which 53% occurred during the
first four days of admission, suggesting that children had medical complications (also linked
with high HIV prevalence). Nevertheless, mortality was also lower in the group with
kwashiorkor (16%) compared to marasmic (47%) and marasmic kwashiorkor (42%). The
comparative mortality rates in children with marasmus, kwashiorkor or marasmic
kwashiorkor treated in outpatient care were reported in four studies (6,18-20) (Table 5). The
mortality rates for children with kwashiorkor ranged from 0% to 4% and were consistently
lower than those with marasmus (range from 4% to 5%) or marasmic kwashiorkor (range
from 7% to 14%).

The comparison of the mortality rate in severely malnourished children treated in inpatient vs
outpatient care after inpatient stabilization was reported in two studies from Malawi and
Bangladesh (14,15) (Table 4). In Malawi, the mortality rate was 5.4% in inpatient and 3.0%
for home-based therapy with RUTF. In Bangladesh, the mortality rate was 3.5% in inpatient,
5% in day care and 3.5% at home (after one week of day care). The higher mortality rate for
inpatients in the Malawian study could have been associated with a higher incidence of
infections for children who were hospitalized (fever, cough and diarrhoea; Wilcoxon’s signed
rank test, p<0.001 for all comparisons). The proportion of children with oedema in the two
studies varied from 81% (Malawi) to 98% (Bangladesh), but their specific mortality rates
were not reported. However, the mortality rate of a subgroup of children with oedema treated
solely in outpatient care in the Malawian study was 4.0% and none of the children with
oedema died after the transfer to domiciliary care in Bangladesh (personal communication of
the author).

3.2.3 Weight/length/MUAC gain and recovery time

Five studies reported the mean or median weight gain in children with oedema treated in
outpatient setting. This did not vary much and ranged between 2.7 g/kg/day and 4.0 g/kg/day
(5,6,14,17,19) (Table 3). All those studies, except the studies in Bangladesh and Ethiopia,
calculated weight gain during the first four weeks of treatment and did not specify whether
this weight gain was registered after loss of oedema.

The median time to recover in outpatient care was only reported in two studies and ranged
from 35 to 42 days (IQR 28–45) (14,19). In one study, the median time to recover was
comparable for kwashiorkor and marasmic cases and longer for marasmic kwashiorkor cases
(19) (Table 5). However, the weight gain after loss of oedema was slower in the kwashiorkor
group.

One study in Bangladesh compared the median time to recover in children treated in home
care after one week of day care vs day care vs inpatient care (14). The group treated at home
took significantly longer to lose oedema (respectively, 19, 13 and 11 days) and to recover
(respectively, 35, 23 and 18 days) and the weight gain after loss of oedema was slower than in
the other two groups (Table 4). It is important to highlight that there was no nutritional
support for the group of children treated at home in this study. On the contrary, in the
Malawian study (80% oedema) (15) the weight gain during the eight weeks of treatment was
higher in the outpatient group treated with RUTF (3.7 g/kg/day) compared to the inpatient
group (3.0 g/kg/day) (Table 4). This study calculated weight gain during the first four weeks
of treatment (total duration of eight weeks) and did not mention whether weight gain was
taken after loss of oedema in the subgroup of children with oedematous malnutrition.

Only three studies reported on mean length/height and MUAC gain in children with oedema
treated in outpatient care (5,6,17) (Table 3). These were consistent and ranged for
length/height gain between 0.24 and 0.39 mm/day and for MUAC gain between 0.2 and 0.3
mm/day. The rate of MUAC and length gain was higher in the outpatient group (with

9
distribution of RUTF) compared to the inpatients in one study with 80% oedema (15) (Table
4).

3.2.4 Costs
Only one study explored the cost per child for SAM treated in inpatient care vs day care vs
home care after one week of day care and found that the latter was four times more cost
effective than inpatient care (13) (Table 4). However, the parental costs were highest for
domiciliary care as no food supplements were provided. No studies explored the cost per
child for the specific group of children with oedema.

3.2.5 Adverse events

Default rates
The default rates for children with oedema treated in outpatient care were reported in six
studies and varied between 3.6% and 15% (Table 3). In the study from Malawi with 15%
drop-out, children were first treated in a tertiary hospital (18). Another study in Bangladesh
found the lowest defaulter rate in inpatient care (1%), followed by outpatient care (4%) and
day care (17%) (14) (Table 4). In the group of children treated in outpatient care, all
defaulters occurred during the first week of day care (personal communication of the authors).
Defaulting appeared similar in the group of children with kwashiorkor (range from 1% to
15%) compared to marasmic children (range from 6% to 19%) (four studies; Table 5).

Infection rate
One study explored the prevalence of fever, cough and diarrhoea as secondary outcomes (15).
The group of children who were managed in inpatient setting had twice the incidence of
symptoms of infection during their recovery periods than the children in home-based
treatment with RUTF. This could have influenced the outcome of inpatient vs outpatient
treatment in this study and contributed to the higher mortality rate and lower recovery rate in
the group of inpatients.

3.3 Quality appraisal

Table 6 summarizes the quality appraisal for the selected studies. Overall, the quality of the
evidence was low. The definition of oedema was poorly or not defined in five of the eight
studies. The precision and accuracy of anthropometric measurement was not addressed in
most of the studies, which may influence the reliability of outcome measurements (recovery,
weight/height/MUAC gain). Weight gain was sometimes measured during the first four
weeks of treatment and it was not mentioned whether assessment was done after resolution of
oedema (5,6,17).

4. Discussion
This systematic review did not retrieve direct evidence on the effectiveness of managing
children >6 months with uncomplicated oedematous malnutrition grade +/++ in the outpatient
vs inpatient setting as no study with a design appropriate to address such a question was
retrieved. However, eight studies reported indirect evidence either by a comparison of
outcomes in oedema cases treated in outpatient care after one week of institutionalized
rehabilitation vs inpatient care, or case series of oedema cases treated in outpatient.

Among the uncomplicated cases of oedematous malnourished children treated exclusively in

outpatient setting, recovery rates exceeded 87% and case fatality rates remained below 4%.
The default rates remained below 10%, indicating that outpatient management is an
acceptable option for the caretakers. These figures conform with the Sphere Guidelines and
the Prudhon index for case fatality rate (11,12). One study exceeded the Sphere standards, but

10
data suggested that there were many complicated cases with high case fatality rates (also
linked with high HIV prevalence) during the first four days of inpatient stabilization (18). It
is, however, quite difficult to disentangle from the reported results what part of this higher
mortality can be attributed to either more severe cases at entry than in other studies, cross-
infection during hospitalization or suboptimal inpatient care. Also in the study in Bangladesh
the weight gain in children with oedema treated in outpatient care did not meet the minimum
Sphere standards (14). The authors of that study suggested that this was due to the absence of
food supplementation in the outpatient treatment group. It needs, however, to be highlighted
that the procedure to assess weight gain was not uniform over the studies and that Sphere
standards are not necessarily adapted to the specific group of children with oedema neither for
outpatient treatment.

Only two studies allowed the comparison of outcomes for outpatient care vs inpatient care
(14,15). It is worth mentioning that the outpatient care occurred after one week of inpatient
stabilization in both studies. In Ciliberto et al. results were much better in outpatient care than
in inpatient care, and the higher infection incidence during rehabilitation in the former group
was presented by the authors as an explanatory factor (15). In Khanum et al., the
rehabilitation was longer and the recovery rate lower in outpatients than in inpatients, but as
already explained above children in outpatient care did not receive food supplements (14). In
five other studies (Table 6), the mortality rate was lower and the recovery rate higher in
children with kwashiorkor than in children with marasmus or marasmic kwashiorkor.

These are encouraging findings indicating that severely malnourished children with oedema
and no other medical complications can be effectively treated in outpatient care. However,
caution is required before generalizing these results to any case of oedematous malnutrition
for three reasons. First, the severity of oedema was poorly defined in all studies. The only
three studies assessing the severity of oedema did so by measuring the depth of the thumb
imprint on the dorsum of the foot, not by referring to the grading +/++/+++. Whether the
good results observed across the studies included in this review relate to populations with a
majority of cases presenting a mild oedema (grade +) is unknown, but the fact that the worst
results were observed in studies where children underwent a first week of inpatient care might
be an indication in that direction (14,15,18). Conversely, worse outcomes in more severe
oedema cases could have been masked by the reporting of mean values as none of the studies
stratified their results by grade of oedema at entry. Second, differences in nutritional status
between defaulters and remainders were not described in the majority of the studies. It is
possible that defaulters were in worse condition than remainders, as suggested in two studies
(5,16), and that their outcomes were also worse. Although the default rates were relatively
low in most studies, these figures may add to death rates. Moreover, most studies were
observational, and the presence of selection bias was difficult to assess, particularly in the
retrospective case series (19,20). It is striking that death rates were higher in the three trials
(14-16) than in the observational studies, with the exception of the study by Sadler et al. (18).
Lastly, six of the eight studies had been carried out in Malawi where the prevalence of HIV is
high and where the ratio kwashiorkor/marasmus in SAM children is potentially higher than in
other settings. The external validity of the results is thus unknown, as the patho-physiology of
oedematous malnutrition is still poorly understood and may vary in different regions.

There were also variations in the management procedures between studies. International
guidelines recommend administration of antibiotics as part of the case management of SAM
cases in outpatient (10). However, the systematic prescription of antibiotics occurred only in
four of eight included studies (14,15,18,19). Nevertheless, this did not seem to have an impact
on the mortality and recovery rates. This coincides with the findings of a retrospective cohort
study in which uncomplicated oedematous cases treated with amoxicillin did not have better
outcomes for mortality and recovery than children who did not have routine antibiotics (20).

11
No cost-effectiveness evaluation was included in our studies for the specific group of children
with oedema treated in outpatient. However, a study in Bangladesh found that home-based
treatment was four times lower than centre-based care and was preferred by mothers (14).
This result related at least partially to the fact that no food supplementation was provided to
the outpatient group in that study. Another review highlighted the scarcity of experimental
evidence on cost-effectiveness of CMAM and that more funding should be invested in
rigorous evaluation of that parameter (21).

In conclusion, our review showed that oedematous malnutrition can be treated effectively in
outpatient service. However, the quality of evidence was low and further good quality studies
in various settings are required before conclusive guidance can be generated. Outpatient
treatment of nutritional oedema already occurs in many programmes. A thorough evaluation
of these programs would complement this review. Outcomes stratified by degree of
nutritional oedema (+/++/+++) are particularly needed. This review also found that marasmic
kwashiorkor patients had a higher risk of death and complications. These patients require
intensive care, monitoring and cautious F-75 based feeding regimens of phase one therapeutic
care that are not possible in outpatient treatment (3). This also emphasizes the need to
combine indicators of SAM, in contradiction with the current WHO guidance of using WHZ,
MUAC or bilateral pitting oedema independently (1). Future RCT should also include
assessment of cost-effectiveness and coverage of the intervention.

12
Table 2
Characteristics of included studies reporting on outpatient treatment of oedematous malnutrition
Prevalence Oedema
Period No. Prevalence of oedema and Definition Admission direct Recovery
Reference Country Intervention Type of study
(month) SAM oedema (%) not wasteda of oedema criteria outpatient criteria
(%) (%)
6–60 months, W/H
Amthor 2009 Non-severe <70% and/or W/H ≥85%
Malawi 5 826 99 71 Outpatient, RUTFb 100 O
(17) <1 cm pitting oedema and (8 weeks)
appetite
10–60 months, Second phase in
W/H >-2 (8
“mild” oedema inpatient therapy Non-randomized
Ciliberto 2005 weeks) and
Mild <0.5 cm and/or W/H <-3 SD (n=113) vs home- (stepped wedge
(15); Ciliberto Malawi 7 645 81 NR 39 oedema
pitting and appetite (no based RUTFb after 1 designd) controlled
2006 (6)c resolved (4
complications if week hospitalization comparative trial
weeks)
direct outpatient) (n= 532)
Outpatient, RUTF 4.2
6–120 months, W/H ≥85%
Present or MJ/day and 24.3 g
Collins 2002 (19) Ethiopia 3.5 170 38 30 W/H ≤70% or 100 and no O retrospective
absent protein/d and Famix
oedema oedema
8.26 MJ/d

Inpatient (n=100) vs
day care (n=200) vs
Khanum 1994 12–60 months, RCT with
Present or home care after 1
(14); Ashworth Bangladesh 12 473 98e 15 oedema and/or W/H 0 W/H >80% sequential
absent week day care (n=173)
1996 (13) <60% allocationf
No food
supplementation in
home care
6–60 months, Oedema
Mild and
Linneman 2007 SAM: oedema resolved
Malawi 12 2131 98 NR moderate <1 Outpatient, RUTFb 77g O
(5) and/or W/H < 70% and W/H
cm pitting
and appetite >85%
6–59 months, W/H >-2
Outpatient, RUTF
Present or SAM: oedema SD and no
Oakley 2010 (16) Malawi 10 1874 78 67 25% milk (n=945) vs 100 RCT, double blind
absent and/or W/H <-3SD oedema 8
15% milk (n=929)
and appetite weeks
Home-based RUTF
6–60 months, W/H
Present or 170 kcal/kg/d + family
Sadler 2008 (18) Malawi 12 1044 78 68 <70% and/or 0 W/H >85% O
absent ration 700 kcal/d after
oedema
inpatient stabilization
6–59 months, W/H >-2
Outpatient, RUTFb
SAM: oedema SD and no
Present or with 7 d amoxy
Trehan 2010 (20) Malawi 24 2453 80 NR and/or W/H <-3 SD 100 oedema (4 O retrospective
absent (n=498) vs no amoxy
and appetite, weeks and
(n=1955)
uncomplicated 12 weeks)
No. = number; NR = not reported; O = observational
a
Wasted defined by anthropometric measurement W/H <-3 SD, except in Ciliberto 2006 W/H <-2 SD.
b
RUTF 175 kcal/kg/d and 5.3 g protein/kg/d for 8 weeks.
c
75% overlap in population with oedema treated exclusively in outpatient (personal note from author) (n=171 vs n=219).
d
First three weeks of centre participation; only inpatient.
e
80% still had oedema at start of outpatient (personal note from author).

13
f
Allocation to treatment group through daily rotation.
g
Proportion of SAM following 8 weeks; outpatient RUTF.

Table 3
Outcomes of severely malnourished child with oedema treated in outpatient care
Median time to Mean length Mean MUAC
No. Mortality No Predicted No. Recovery Mean weight gain Use of resources Other adverse
Reference recover (d) gain (mm/d) +/- gain (mm/d) +/-
oedema (%) death death (12) (%) (g/kg/d) +/- SD cost/child effects
(IQR) SD SD

Amthor
Default 3.6%,
Malawi 2009 819 0.9 8 8 93.7 NR 2.7±3.7 0.3±0.9 0.2±0.3 NR
failure 1.8%
(17)a

Ciliberto
Overall default
Malawi 2005 171 4b 8 30 89b NR 2.8±3.2b 0.24±0.27b 0.3±0.4b NR
9.6%, relapse 3%
(6,15)
Collins
Transferred 0%,
Ethiopia 2002 50 0 0 NR 96 42 (28–45) 2.7 (0.0–.8)d NR NR NR
default 4%
(19)c
Khanum 1994;
Institutional cost
Ashworth 1997 Default 4%, late
170 3.5 6 NR 75 35 (NR) 4±NR NR NR $29/child, parental
Bangladesh exclusionf 17%
cost $9/child
(13,14)e
Linneman
Default 7%, failure
Malawi 2007 2090 1 29 64 89 NA 3.5±4.1 0.39±0.54 0.24±0.35 NR
3%
(5)g
Oakley
Malawi 2010 1458 3.5 64 58 88 NR NR NR NR NR Default 3%
(16)h
Sadler Malawi
731 16 117 NR 69 NR NR NR NR NR Default 15%
2008 (18)

Trehan Malawi Default 8.3 %,

1962 1.5 29 NR 87.7 NR NR NR NR NR
2010 (20) failure 2.5

No. = number; NR = not reported; $ = United States dollar

a
Outcomes of SAM (99% oedema).
b
Outcomes for the group presenting oedema (kwashiorkor and marasmic kwashiorkor) and treated exclusive outpatient. Outcomes group oedema and no wasting in OTP (Ciliberto 2006): recovery
95%, mortality 4%, default 1%, failure 1%.
c
Outcomes group oedema and no wasting in OTP.
d
Median weight gain (IQR).
e
Outcomes of SAM (98.5% oedema) in outpatient. 80% had oedema at start of home care; mortality 0%, recovery 100% in oedema and non-oedema cases.
f
Due to tuberculosis or blood transfusion.
g
Outcomes of SAM with oedema (98%) exclusively treated in outpatient (77%) were not reported separately.
h
Mortality (4%), defaulting (3%), failure (8%) and transfer (2%) only reported for the global group (kwashiorkor or marasmus or both).

14
Table 4
Comparison of outcomes for different treatment approaches
Predicte Weight Length MUAC Cost/child
Time to Prevalence Prevalence Prevalence Relapse
Mortality No. d no. Recovery gain gain gain ($a) Relapse Default
Reference Treatment recover fever ± SD cough ± SD diarrhoea ± (6month
(%) deaths deaths (%) (g/kg/d) ± (mm/d) ± (mm/d) ± institutional/ (%) rate
(d) (d) (d) SD (d) s)b (%)
(12) SD SD SD parental
Inpatient 6.2 10 9 49 NR 3.0±8.8 0.04±0.35 0.28±0.44 NA 1.8±3.3c 1.8±3.6c 1.3±2.7c 10.6 39.1 8.1b
Ciliberto
2005 (15) Outpatient (1
3.7 30 30 72 NR 3.7±4.3 0.20±0.33 0.42±0.71 NA 1.0±2.0c 0.8±2.4c 0.7±1.7c 6.2 2.2 9.8b
inpatient)
Inpatient 3.5 7 NR 86.5 18 11 NR NR 155.9/3.1 NR NR NR NR NR 1.0
Khanum
1994; (14) Day care 5.0 10 NR 67.0 23 6 NR NR 59.3/4.5 NR NR NR NR NR 17.0
Ashworth
1997 (13) Outpatient (1
3.5 6 NR 75.1 35 4 NR NR 29.4/9.4 NR NR NR NR NR 4.0
day care)

No. = number; NR not reported

a
United States dollar.
b
Default rates in global study group of MAM and SAM.
c
Wilcoxon’s signed-rank test, p<0.001 for all comparisons.

15
 Table 5
Comparison of outcomes of outpatient treatment in kwashiorkor – marasmic kwashiorkor – marasmus

Reference Mortality (%) Recovery (%) Median time to recover (d) (IQR) Weight gain (g/kg/d) +-SD Defaulter (%) Transferred (%)

M K MK M K MK M K MK M K MK G M K MK M K MK
Ciliberto 2006 (6) NR 4 7 NA 95 65 NR NR NR NR NR NR 4.0 NR 1 8 NR NR NR
a a a
Collins 2002 (19) 5 0 14 81 96 71 42 (28–56) 42 (28–45) 56 (42–70) 4.8 (2.9–8.1) 2.7 (0.0–4.8) 4.0 (2.7–4.3) 5 6 4 0 9 0 14
b
Oakley 2010 (16) NR NR NR 70 88 NR NR NR NR NR NR NR 3 NR NR NR NR NR NR
Trehan 2010 (20) 3.7 1.5 NR 75.1 87.7 NR NR NR NR NR NR NR 9 11.8 8.3 NR NR NR NR
Sadler 2008 (18) 47 16 42 33 69 38 NR NR NR NR NR NR 16.2 19 15 20 NR NR NR

M = marasmus, K = kwashiorkor; MK = marasmic kwashiorkor defined as oedema and wasted W/H <-3 SD, except for Ciliberto 2006 where wasted is defined as W/H <-2 SD; G = global malnutrition (M or K or
MK); NR = not reported
a
Median weight gain (IQR).
b
Mortality (4%), defaulting (3%), failure (8%), weight gain (2.2 mg/kg/d), MUAC gain (0.15 mm/d), and transfer (2%) only reported for the global group.

16
Table 6
Quality appraisal studies reporting on outcome of children with oedema
Sources +
methods
of Statistical
Reference Reliability of measures Missing data Other comments
selection analysis
correctly
described
Oedema
Training Quality Refere % Differences of
Definition Double assessed
of control nce drop- drop-outs
of oedema measurement in all
assessors (method) testa out assessed
children
P (median test,
Amthor 2009 M A A NR NR NR NR 3.6 NR Case series, Malawi (high HIV prevalence)
p)
A, intention to
treat analysis, Non-randomized controlled trial, Malawi,
Ciliberto 2005; 2006 M NR A NR NR NR A 9.6 NR
difference in CI high HIV prevalence
95%
P, (χ2 for Retrospective, no inpatient facility available
Collins and Sadler 2002 M A NR NR A NR A 5 NR
proportions) All oedema cases treated outpatient
RCT with sequential allocation
Khanum 1994; Ashworth 1997 M NR NR NR NR NR NR NR 9.8 NR
Main focus cost-effectiveness analysis
Drop-outs had
lower
Linneman 2007 M A A NR NR NR NR A 7.4 Case series, Malawi (high HIV prevalence)
anthropometri
c indices
A, Intention to Drop-outs
treatment, Cox were younger (RCT double blind, two types RUTF),
Oakley 2010 M A NR A A A NR 3
regression and more Malawi
(M,K) marasmatic
Sadler 2008 P A NR NR A NR NR NR 15 A Case series, initial inpatient, high HIV %
A (for the
assessment of the retrospective, uncomplicated SAM and
Trehan 2010 P NR NR NR NR NR differences due to
NR 8.3 NR
RUTF, all oedema cases treated outpatient
antibiotic)

NR = not reported; M = moderate; A = appropriate; P = poor

a
Results of index test were interpreted relative to the reference standard.

17
Annexes
Annex 1: Overview of 33 guidelines on CMAM
Country Year Oedema +/++ in outpatient?
Afghanistan 2008 Yes
Angola 2008 No
Bangladesh 2010 Yes
Botswana 2009 Yes
Burundi 2010 Yes
Central African Republic unknown Yes
Democratic Republic of the Congo 2008 Yes
Djibouti 2009 Yes
Ethiopia 2007 Yes
Eritrea Extract with admission/discharge criteria - 2010 Yes
Ghana 2010 Yes
Honduras Revisions without date Yes
Côte Ivoire 2010 Only oedema +
Kenya 2008 Only oedema +
Madagascar 2009 Yes
Malawi Draft without date Yes
Mali 2007 No
Mauretania 2007 Yes
Mozambique 2010 No
Republic of Niger 2009 No
Rwanda No date Yes
Sierra Leone 2009 Yes
Somalia 2010 Yes
Republic of the Sudan 2009 Yes
Republic of South Sudan 2009 Yes
Sri Lanka 2007 Yes
Tanzania 2010 No
Tajikistan 2009 Yes
Togo 2009 Yes
Uganda 2010 Yes
Yemen 2008 Yes
Zanzibar 2010 No
Zimbabwe 2008 Yes

18
Annex 2
Key words used for database other than Medline

The key words used in:

 Cochrane library: “nutritional edema”
 Embase: “(((‘kwashiorkor’/exp OR (‘malnutrition’/exp AND ‘edema’/exp) OR (‘nutritional
deficiency’/exp AND ‘edema’/exp) OR (wast* AND ‘edema’/exp)) AND (‘infant’/exp OR
‘child’/exp)) AND (‘community care’/exp OR treatment OR ‘rehabilitation’/exp OR
‘management’/exp OR ‘hospitalization’/exp)) NOT (‘editorial’/exp OR ‘letter’/exp OR comment)) ”,
 DARE: “kwashiorkor OR (nutrition AND edema)”

19
Annex 3
Studies excluded and reasons for exclusion

Review examining the effectiveness of 16 studies with community-based rehabilitation for treatment of SAM.
1. Ashworth 2006 (22)
Rehabilitation is referring to the second phase of treatment. No reporting on prevalence of oedema.
2. Birem-Etchebes 1988 (23) Intervention: All kwashiorkor and marasmic cases were hospitalized. Qualitative study. No data on main outcomes.
45 children with severe, moderate or mild malnutrition (following the Gomez classification) followed in outpatient.
3. Bredow 1994 (24)
Only four (9%) oedema and no reporting of outcomes of these children with oedema.
Review on community-based management of SAM and moderate acute malnutrition (MAM). No reporting on
4. Brown et al. 2009 (25)
prevalence of oedema.
5. Chaiken 2006 (26) No reporting of prevalence and outcomes of patients with oedema in outpatient.
6. Chapko 1994 (27) RCT: comparison of hospital vs day care. 24% oedema. No reporting of outcomes in patients with oedema.
7. Collins 2006 (3) Review on community based-management of SAM with RUTF. No reporting on prevalence of oedema.
8. Diop 2003 (28) Inclusion after oedema disappeared.
No reporting on prevalence of severely malnourished children with oedema, only 12 cases of third degree
9. Fernandez-Concha 1991 (29)
malnutrition (W/A).
Comparison of inpatient, inpatient followed by home treatment and exclusive home treatment. Cases of oedema
10. Gaboulaud 2007 (8)
were excluded from group with exclusive home treatment.
11. Gossens 2012 (30) Children with oedema were treated in in-patient setting
12. Gueri 1985 (31) No reporting on prevalence oedema, no data on main outcomes.
13. Heikens 1989 (32) Exclusion of oedema cases in outpatient.
14. Husaini 1982 (33) Very small sample (five children with kwashiorkor).
15. Jansen 1986 (34) Supplementary feeding at home. No reporting on prevalence oedema, no data on main outcomes.
16. Kumari 1975 (35) No data on main outcomes.
Moderately and severely malnourished children were treated in inpatient (complications), outpatient or inpatient
17. Lapidus 2009 (36) followed by outpatient. Risk factor analysis for mortality. No data on prevalence and outcomes of patients with
oedema in outpatient.
Intervention: Testing of three dietary regimens, all obtained initial inpatient stabilization (>80% oedema during
18. Manary 2004 (37)
hospitalization). No reporting of prevalence and outcomes of patients with oedema during outpatient phase.
19. Manary 2008 (38) Review on management of MAM and SAM. No reporting on outcomes.
Idem as Manary 2004, but with HIV positive children. All obtained inpatient stabilization (<80% oedema). No
20. Ndekha 2004 (39)
reporting of prevalence and outcomes of patients with oedema in outpatient.
21. Prudhon 1997 (12) Model to calculate the number of deaths based on nutritional status (including presence of oedema).
Intervention: All obtained inpatient stabilization. 12% received exclusive outpatient. Percentage oedema and
22. Sandige 2004 (7)
outcomes during outpatient phase is not reported.
Evaluation of community-based treatment model of SAM. Prevalence of oedema 61%. Comparison between SAM
with complications (78%) treated first in inpatient and those without complications treated indirectly in outpatient.
23. Sadler (40)
Different triage criteria: Children with oedema ++/+++ were admitted in inpatient, children with oedema + in
outpatient. The outcomes for children with oedema +/++/+++ were reported separately, but inconsistent with
20
 Collins 2006.
24. Schoffield 1996 (41) Review on causes of mortality in treatment of SAM.
25. Shah 1971 (42) Case report of 16 cases of Kwashiorkor. No data on main outcomes.
26. Tellier 1996 (43) No data on outcomes for SAM with oedema treated in outpatient.

21
References
1. A Joint Statement by the World Health Organization and United Nations Children's Fund. WHO child growth
standards and the identification of severe acute malnutrition in infants and children. Geneva, WHO/UNICEF,
2009.

2. Beattie J, Herbert PH, Bell DJ. Famine oedema. British Journal of Nutrition, 1948, 2:47–65.

3. Collins S et al. Management of severe acute malnutrition in children. Lancet, 2006, 368:1992–2000.

4. Management of severe malnutrition: a manual for physisians and other senior health workers. Geneva,
WHO, 1999.

5. Linneman Z et al. A large-scale operational study of home-based therapy with ready-to-use therapeutic food
in childhood malnutrition in Malawi. Maternal & Child Nutrition, 2007, 3:206–215.

6. Ciliberto MA et al. Home-based therapy for oedematous malnutrition with ready-to-use therapeutic food. Acta
Paediatrica, 2006, 95:1012–1015.

7. Sandige H et al. Home-based treatment of malnourished Malawian children with locally produced or imported
ready-to-use food. Journal of Pediatric Gastroenterology and Nutrition, 2004, 39:141–146.

8. Gaboulaud V et al. Could nutritional rehabilitation at home complement or replace centre-based therapeutic
feeding programmes for severe malnutrition? Journal of Tropical Pediatrics, 2007, 53:49–51.

9. Collins S et al. Key issues in the success of community-based management of severe malnutrition. Food and
Nutrition Bulletin, 2006, 27:S49–S82.

10. A Joint Statement by the World Health Organization, the World Food Programme, the United Nations System
Standing Committee on Nutrition and the United Nations Children's Fund. Community-based management of
severe acute malnutrition. Geneva, WHO/WFP/UN/UNICEF, 2007.

11. The Sphere Handbook: Humanitarian Charter and Minimum Standards in Humanitarian Response. Geneva,
The Sphere Project, 2011.

12. Prudhon C et al. A model to standardise mortality of severely malnourished children using nutritional status
on admission to therapeutic feeding centres. European Journal of Clinical Nutrition, 1997, 51:771–777.

13. Ashworth A, Khanum S. Cost-effective treatment for severely malnourished children: what is the best
approach? Health Policy and Planning, 1997, 12:115–121.

14. Khanum S, Ashworth A, Huttly SR. Controlled trial of three approaches to the treatment of severe
malnutrition. Lancet, 1994, 344:1728–1732.

15. Ciliberto MA et al. Comparison of home-based therapy with ready-to-use therapeutic food with standard
therapy in the treatment of malnourished Malawian children: a controlled, clinical effectiveness trial.
American Journal of Clinical Nutrition, 2005, 81:864–870.

16. Oakley E et al. A ready-to-use therapeutic food containing 10% milk is less effective than one with 25% milk
in the treatment of severely malnourished children. Journal of Nutrition, 2010, 140:2248–2252.

17. Amthor RE, Cole SM, Manary MJ. The use of home-based therapy with ready-to-use therapeutic food to treat
malnutrition in a rural area during a food crisis. Journal of the American Dietetic Association, 2009, 109:464–
467.

18. Sadler K et al. Improving the management of severe acute malnutrition in an area of high HIV prevalence.
Journal of Tropical Pediatrics, 2008, 54:364–369.

19. Collins S, Sadler K. Outpatient care for severely malnourished children in emergency relief programmes: a
retrospective cohort study. Lancet, 2002, 360:1824–1830.
22
20. Trehan I et al. Evaluation of the routine use of amoxicillin as part of the home-based treatment of severe acute
malnutrition. Tropical Medicine & International Health, 2010, 15:1022–1028.

21. Bachmann MO. Cost-effectiveness of community-based treatment of severe acute malnutrition in children.
Expert Review of Pharmacoeconomics & Outcomes Research, 2010, 10:605–612.

22. Ashworth A. Efficacy and effectiveness of community-based treatment of severe malnutrition. Food and
Nutrition Bulletin, 2006, 27:S24–S48.

23. Birem-Etchebes S, Gonzalez JP. Use of local food resources in the rehabilitation of children with malnutrition
in the Central African Republic. Bulletin de la Societe de Pathologie Exotique et de ses Filiales, 1988,
81:769–776.

24. Bredow MT, Jackson AA. Community-based, effective, low-cost approach to the treatment of severe
malnutrition in rural Jamaica. Archives of Disease in Childhood, 1994, 71:297–303.

25. Brown KH, Nyirandutiye DH, Jungjohann S. Management of children with acute malnutrition in resource-
poor settings. Nature Reviews Endocrinology, 2009, 5:597–603.

26. Chaiken MS, Deconinck H, Degefie T. The promise of a community-based approach to managing severe
malnutrition: a case study from Ethiopia. Food and Nutrition Bulletin, 2006, 27:95–104.

27. Chapko MK et al. Randomized clinical trial comparing hospital to ambulatory rehabilitation of malnourished
children in Niger. Journal of Tropical Pediatrics, 1994, 40:225–230.

28. Diop el HI et al. Comparison of the efficacy of a solid ready-to-use food and a liquid, milk-based diet for the
rehabilitation of severely malnourished children: a randomized trial. American Journal of Clinical Nutrition,
2003, 78:302–307.

29. Fernandez-Concha D, Gilman RH, Gilman JB. A home nutritional rehabilitation programme in a Peruvian
peri-urban shanty town (pueblo joven). Transactions of the Royal Society of Tropical Medicine and Hygiene,
1991, 85:809–813.

30. Goossens S et al. Mid-upper arm circumference based nutrition programming: evidence for a new
approach in regions with high burden of acute malnutrition. PLoS One, 2012, 7:e49320.

31. Gueri M et al. A supplementary feeding programme for the management of severe and moderate malnutrition
outside hospital. Journal of Tropical Pediatrics, 1985, 31:101–108.

32. Heikens GT et al. The Kingston Project. I. Growth of malnourished children during rehabilitation in the
community, given a high-energy supplement. European Journal of Clinical Nutrition, 1989, 43:145–60.

33. Husaini YK et al. Outpatient rehabilitation of severe protein energy malnutrition. Food and Nutrition Bulletin,
1982.

34. Jansen AA, Verkley MT. Ambulatory home nutrition rehabilitation in rural Kenya. Journal of Tropical
Pediatrics, 1986, 32:258–262.

35. Kumari VK, Damodaran M, Rau P. Evaluation of domicilliary management of protein-calorie malnutrition.
Tropical and Geographic Medicine, 1975, 27:99–102.

36. Lapidus N et al. Mortality risk among children admitted in a large-scale nutritional program in Niger, 2006.
PLoS One, 2009, 4:e4313.

37. Manary MJ et al. Home-based therapy for severe malnutrition with ready-to-use food. Archives of Disease in
Childhood, 2004, 89:557–561.

38. Manary MJ, Sandige HL. Management of acute moderate and severe childhood malnutrition. British Medical
Journal, 2008, 337:a2180.

23
39. Ndekha MJ et al. Home-based therapy with ready-to-use therapeutic food is of benefit to malnourished, HIV-
infected Malawian children. Acta Paediatrica, 2005, 94:222–225.

40. Sadler, K. Community-based therapeutic care: treating severe acute malnutrition in sub-Saharan Africa
[thesis]. London, University College London, 2009.

41. Schofield C, Ashworth A. Why have mortality rates for severe malnutrition remained so high? Bulletin of the
World Health Organization, 1996, 74:223–229.

42. Shah PM, Udani PM, Aphale RV. Domiciliary management of kwashiorkor in rural set up: a longitudinal
study of clinical, economic and social aspects. Indian Pediatrics, 1971, 8:805–813.

43. Tellier V et al. Ambulatory health care of moderately and severely malnourished children in the rural health
zone of Kapolowe in Shaba, Zaire. Sante, 1996, 6:213–219.

24